In this project we answer the question: Does the amount of screen time a person spends at age 16 affect their levels of depression and anxiety at age 18?
In order to answer this, we applied a range of machine learning methods to a dataset which was synthesised from a subset of the ALSPAC (Avon Longitudinal Study of Parents and Children) cohort study.
The machine learning algorithms implemented were: Logistic Regression, K Nearest Neighbours Classification, Support Vector Machine Classification, XGBoost and a Forward Neural Network.
Overall, none of the models were able to predict depression based on the screen time any better than random choice when evaluated on data resembling the real world. Although we did achieve higher scores when training and evaluating on a balanced dataset.